AI against asymptomatic COVID spread.

By
2021-02-19
4 mins read

Researchers at the Massachusetts Institute of Technology have developed an artificial intelligence model that can distinguish between a healthy cough and one that comes from an asymptomatic coronavirus patient with 98.5% accuracy.

 

When the pandemic had hit, Subirana’s research team at MIT had been working on a set of machine learning algorithms to detect Alzheimer’s disease in audio recordings using biomarkers such as vocal cord strength, sentiment, lung performance, and muscular degradation. When it became clear that coughing was a key feature of COVID-19, they quickly pivoted to see if their model could detect coronavirus infections.

 

“The sounds of talking and coughing are both influenced by the vocal cords and surrounding organs. This means that when you talk, part of your talking is like coughing, and vice versa,” Subirana said in the release.

 

“It also means that things we easily derive from fluent speech, AI can pick up simply from coughs, including things like the person’s gender, mother tongue, or even emotional state. There’s in fact sentiment embedded in how you cough.”

The team felt confident that the model would work as there was growing evidence that COVID patients experienced similar symptoms associated with Alzheimer’s such as neuromuscular impairment that affects the vocal cords. In April, they gathered 70,000 recordings of people forcibly coughing into their cell phones or laptops, which amounted to the largest cough dataset that we know of, containing more than 2,00,000 cough audio samples. About 2,500 of those were submitted by people with COVID-19. Participants also had to answer surveys about symptoms they were experiencing, their COVID-19 diagnosis, gender, geographical location and native language.

 

The researchers then trained the model on tens of thousands of samples of cough, as well as spoken words. When new cough recordings were fed to the model, by listening for vocal cord strength, sentiment, lung and respiratory performance, and muscular degradation, it identified 98.5% of coughs from coronavirus patients, and 100% of asymptomatic coughs.

 

The team said that it is working on incorporating the model into apps and eventually, smart speakers and other listening devices so that people can consistently and conveniently be screened for coronavirus infection. That can help prevent asymptomatic individuals from unknowingly spreading the virus to others and hence, help us fight the war against COVID19. Users could cough into their phones daily and could instantly know if they might be infected and therefore should confirm with a formal test or not.

 

Imagine, coughing up your phone daily and asking your Google assistant if you should go and get a COVID test or not. That might sound a little eccentric. But if it works, is free and can save you from having a cotton swab poked up your nose, then why not.

 

Tags:

techblog

medicalScience

Covid19

ComputerScience

Artificial Intelligence

ai